BP神经网络隐层单元数的确定方法及实例

针对BP神经网络隐层单元数不易确定的问题,提出一种在传统的经验公式基础上快速确定隐层单元数的方法。该方法首先借助经验公式确定隐层单元数的取值范围,然后将其扩大,在这个扩大的范围内寻找最优值。以BP神经网络预测交通流量为例,解释说明了具体的步骤,以及网络模型的隐层结构对模型仿真精度的影响。结果表明,采用该方法可快速决定隐层单元数,在实例中采用16个隐层单元数为最佳。     

二级倒立摆的Sugeno型模糊神经网络控制

为了提高二级倒立摆系统实时控制的响应速度和稳定性,在设计Mamdani型模糊推理规则控制器控制倒立摆系统稳定的基础上,设计了一种更有效率的基于Sugeno型模糊推理规则的模糊神经网络控制器.该控制器使用BP神经网络和最小二乘法的混合算法进行参数训练.能够准确归纳输入输出量的模糊隶属度函数和模糊逻辑规则.通过与Mamdani型控制器的仿真对比及实际控制实验结果,表明该Sugeno型模糊神经网络控制器时二级倒立摆实验装置的控制具有良好的稳定性、快速性和较高的控制精度.     

基于混沌时间序列的变电站设备温度预测

变电站输变线路和设备的温度变化能够反映其老化、负载过高等引起的安全隐患.通过对变电站设备温度数据的非线性分析和预测,实现对设备的有效预警,将避免事故引起的巨大损失.对变电站已测温度数据建立时间序列,利用小数据量法验证变电站设备温度时间序列的混沌特性.研究基于RBF神经网络的混沌时间序列预测并与神经网络预测进行对比,单步预测与多步预测结果均优于神经网络预测.仿真结论证明了基于神经网络的混沌时间序列预测方法的有效性.     

永磁同步电机的自适应反演滑模变结构控制

针对永磁同步电机提出一种基于反演的PMSM自适应滑模控制方案.设计基于反演的滑模变结构位置控制器,通过RBF神经网络实现系统参数变化和外部负载扰动等引起的不确定上界值的在线辨识,减小滑模控制器的控制量,并引入饱和函数来减弱系统的"抖动"现象.理论分析和仿真结果对比表明,基于RBF神经网络的自适应反演滑模控制对参数变化和外部负载扰动具有很好的鲁棒性,永磁同步电动机获得了很好的跟踪效果.     

A method for diagnosis of current faults in antenna arrays using neural networks

A method for detection of faulty elements in antenna arrays from far‐field radiation pattern is presented. The proposed technique finds variation of current from correct values in the faulty elements. A step wise approach is proposed to determine magnitude and phase of current excitation and location of faulty element using neural networks. The results with radial basis function neural network and probabilistic neural network are compared. © 2008 Wiley Periodicals, Inc. Int J RF and Microwave CAE, 2009.     

Boundedness and Convergence of Online Gradient Method with Penalty for Linear Output Feedforward Neural Networks

This paper investigates an online gradient method with penalty for training feedforward neural networks with linear output. A usual penalty is considered, which is a term proportional to the norm of the weights. The main contribution of this paper is to theoretically prove the boundedness of the weights in the network training process. This boundedness is then used to prove an almost sure convergence of the algorithm to the zero set of the gradient of the error function.     

Membrane computing and brain modelling

We briefly discuss variants of (extended) spiking neural P systems that combine features from the areas of membrane computing and spiking neurons. This work was presented in part at the 13th International Symposium on Artificial Life and Robotics, Oita, Japan, January 31–February 2, 2008     

The asymptotic optimization of pre-edited ANN classifier

The generalization problem of an artificial neural network (ANN) classifier with unlimited size of training sample, namely asymptotic optimization in probability, is discussed in this paper. As an improved ANN network model, the pre-edited ANN classifier shows better practical performance than the standard one. However, it has not been widely applied due to the absence of the related theoretical support. To further promote its application in practice, the asymptotic optimization of the pre-edited ANN classifier is studied in this paper. To help study ANN asymptotic optimization in probability, we gives a review of the previous research works on asymptotic optimization in probability of non-parametric classifier, and grouped the main methods into four classes: two-step method, one-step method, generalization method and hypothesis method. In this paper, we adopt generalization/hypothesis mixed method to prove that pre-edited ANN is asymptotically optimal in probability. Furthermore, a simulation is presented to provide an experimental support for our theoretical work.     

A neuro-genetic based short-term forecasting framework for network intrusion prediction system

Information systems are one of the most rapidly changing and vulnerable systems, where security is a major issue. The number of security-breaking attempts originating inside organizations is increasing steadily. Attacks made in this way, usually done by ＂authorized＂ users of the system, cannot be immediately traced. Because the idea of filtering the traffic at the entrance door, by using firewalls and the like, is not completely successful, the use of intrusion detection systems should be considered to increase the defense capacity of an information system. An intrusion detection system （IDS） is usually working in a dynamically changing environment, which forces continuous tuning of the intrusion detection model, in order to maintain sufficient performance. The manual tuning process required by current IDS depends on the system operators in working out the tuning solution and in integrating it into the detection model. Furthermore, an extensive effort is required to tackle the newly evolving attacks and a deep study is necessary to categorize it into the respective classes. To reduce this dependence, an automatically evolving anomaly IDS using neuro-genetic algorithm is presented. The proposed system automatically tunes the detection model on the fly according to the feedback provided by the system operator when false predictions are encountered. The system has been evaluated using the Knowledge Discovery in Databases Conference （KDD 2009） intrusion detection dataset. Genetic paradigm is employed to choose the predominant features, which reveal the occurrence of intrusions. The neuro-genetic IDS （NGIDS） involves calculation of weightage value for each of the categorical attributes so that data of uniform representation can be processed by the neuro-genetic algorithm. In this system unauthorized invasion of a user are identified and newer types of attacks are sensed and classified respectively by the neuro-genetic algorithm. The experimental results obtained in this work show that the system achieves improvement in terms of misclassification cost when compared with conventional IDS. The results of the experiments show that this system can be deployed based on a real network or database environment for effective prediction of both normal attacks and new attacks.     

An improved differential evolution trained neural network scheme for nonlinear system identification

This paper presents an improved nonlinear system identification scheme using di？erential evolution （DE）, neural network （NN） and Levenberg Marquardt algorithm （LM）. With a view to achieve better convergence of NN weights optimization during the training, the DE and LM are used in a combined framework to train the NN. We present the convergence analysis of the DE and demonstrate the efficacy of the proposed improved system identification algorithm by exploiting the combined DE and LM training of the NN and suitably implementing it together with other system identification methods, namely NN and DE＋NN on a number of examples including a practical case study. The identification results obtained through a series of simulation studies of these methods on different nonlinear systems demonstrate that the proposed DE and LM trained NN approach to nonlinear system identification can yield better identification results in terms of time of convergence and less identification error.